regex_primitives.hpp

support vector clustering for vc++

///////////////////////////////////////////////////////////////////////////////
/// \file regex_primitives.hpp
/// Contains the syntax elements for writing static regular expressions.
//
//  Copyright 2004 Eric Niebler. Distributed under the Boost
//  Software License, Version 1.0. (See accompanying file
//  LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)

#ifndef BOOST_XPRESSIVE_REGEX_PRIMITIVES_HPP_EAN_10_04_2005
#define BOOST_XPRESSIVE_REGEX_PRIMITIVES_HPP_EAN_10_04_2005

#include <climits>
#include <boost/mpl/assert.hpp>
#include <boost/preprocessor/cat.hpp>
#include <boost/xpressive/proto/proto.hpp>
#include <boost/xpressive/detail/detail_fwd.hpp>
#include <boost/xpressive/detail/core/icase.hpp>
#include <boost/xpressive/detail/core/action.hpp>
#include <boost/xpressive/detail/core/matchers.hpp>
#include <boost/xpressive/detail/static/as_xpr.hpp>
#include <boost/xpressive/detail/static/compile.hpp>
#include <boost/xpressive/detail/static/modifier.hpp>
#include <boost/xpressive/detail/static/regex_operators.hpp>
#include <boost/xpressive/detail/static/productions/productions.hpp>

namespace boost { namespace xpressive { namespace detail
{

typedef assert_word_placeholder<word_boundary<true> > assert_word_boundary;
typedef assert_word_placeholder<word_begin> assert_word_begin;
typedef assert_word_placeholder<word_end> assert_word_end;

/*
///////////////////////////////////////////////////////////////////////////////
/// INTERNAL ONLY
// BOOST_XPRESSIVE_GLOBAL
//  for defining globals that neither violate the One Definition Rule nor
//  lead to undefined behavior due to global object initialization order.
//#define BOOST_XPRESSIVE_GLOBAL(type, name, init)                                        \
//    namespace detail                                                                    \
//    {                                                                                   \
//        template<int Dummy>                                                             \
//        struct BOOST_PP_CAT(global_pod_, name)                                          \
//        {                                                                               \
//            static type const value;                                                    \
//        private:                                                                        \
//            union type_must_be_pod                                                      \
//            {                                                                           \
//                type t;                                                                 \
//                char ch;                                                                \
//            } u;                                                                        \
//        };                                                                              \
//        template<int Dummy>                                                             \
//        type const BOOST_PP_CAT(global_pod_, name)<Dummy>::value = init;                \
//    }                                                                                   \
//    type const &name = detail::BOOST_PP_CAT(global_pod_, name)<0>::value
*/

} // namespace detail

/// INTERNAL ONLY (for backwards compatibility)
unsigned int const repeat_max = UINT_MAX-1;

///////////////////////////////////////////////////////////////////////////////
/// \brief For infinite repetition of a sub-expression.
///
/// Magic value used with the repeat\<\>() function template
/// to specify an unbounded repeat. Use as: repeat<17, inf>('a').
/// The equivalent in perl is /a{17,}/.
unsigned int const inf = UINT_MAX-1;

/// INTERNAL ONLY (for backwards compatibility)
proto::op_proxy<
    proto::unary_op<detail::epsilon_matcher, proto::noop_tag>
> const epsilon = {};

///////////////////////////////////////////////////////////////////////////////
/// \brief Successfully matches nothing.
///
/// Successfully matches a zero-width sequence. nil always succeeds and
/// never consumes any characters.
proto::op_proxy<
    proto::unary_op<detail::epsilon_matcher, proto::noop_tag>
> const nil = {};

///////////////////////////////////////////////////////////////////////////////
/// \brief Matches an alpha-numeric character.
///
/// The regex traits are used to determine which characters are alpha-numeric.
/// To match any character that is not alpha-numeric, use ~alnum.
///
/// \attention alnum is equivalent to /[[:alnum:]]/ in perl. ~alnum is equivalent
/// to /[[:^alnum:]]/ in perl.
proto::op_proxy<
    proto::unary_op<detail::posix_charset_placeholder, proto::noop_tag>
  , char const *
> const alnum = {"alnum"};

///////////////////////////////////////////////////////////////////////////////
/// \brief Matches an alphabetic character.
///
/// The regex traits are used to determine which characters are alphabetic.
/// To match any character that is not alphabetic, use ~alpha.
///
/// \attention alpha is equivalent to /[[:alpha:]]/ in perl. ~alpha is equivalent
/// to /[[:^alpha:]]/ in perl.
proto::op_proxy<
    proto::unary_op<detail::posix_charset_placeholder, proto::noop_tag>
  , char const *
> const alpha = {"alpha"};

///////////////////////////////////////////////////////////////////////////////
/// \brief Matches a blank (horizonal white-space) character.
///
/// The regex traits are used to determine which characters are blank characters.
/// To match any character that is not blank, use ~blank.
///
/// \attention blank is equivalent to /[[:blank:]]/ in perl. ~blank is equivalent
/// to /[[:^blank:]]/ in perl.
proto::op_proxy<
    proto::unary_op<detail::posix_charset_placeholder, proto::noop_tag>
  , char const *
> const blank = {"blank"};

///////////////////////////////////////////////////////////////////////////////
/// \brief Matches a control character.
///
/// The regex traits are used to determine which characters are control characters.
/// To match any character that is not a control character, use ~cntrl.
///
/// \attention cntrl is equivalent to /[[:cntrl:]]/ in perl. ~cntrl is equivalent
/// to /[[:^cntrl:]]/ in perl.
proto::op_proxy<
    proto::unary_op<detail::posix_charset_placeholder, proto::noop_tag>
  , char const *
> const cntrl = {"cntrl"};

///////////////////////////////////////////////////////////////////////////////
/// \brief Matches a digit character.
///
/// The regex traits are used to determine which characters are digits.
/// To match any character that is not a digit, use ~digit.
///
/// \attention digit is equivalent to /[[:digit:]]/ in perl. ~digit is equivalent
/// to /[[:^digit:]]/ in perl.
proto::op_proxy<
    proto::unary_op<detail::posix_charset_placeholder, proto::noop_tag>
  , char const *
> const digit = {"digit"};

///////////////////////////////////////////////////////////////////////////////
/// \brief Matches a graph character.
///
/// The regex traits are used to determine which characters are graphable.
/// To match any character that is not graphable, use ~graph.
///
/// \attention graph is equivalent to /[[:graph:]]/ in perl. ~graph is equivalent
/// to /[[:^graph:]]/ in perl.
proto::op_proxy<
    proto::unary_op<detail::posix_charset_placeholder, proto::noop_tag>
  , char const *
> const graph = {"graph"};

///////////////////////////////////////////////////////////////////////////////
/// \brief Matches a lower-case character.
///
/// The regex traits are used to determine which characters are lower-case.
/// To match any character that is not a lower-case character, use ~lower.
///
/// \attention lower is equivalent to /[[:lower:]]/ in perl. ~lower is equivalent
/// to /[[:^lower:]]/ in perl.
proto::op_proxy<
    proto::unary_op<detail::posix_charset_placeholder, proto::noop_tag>
  , char const *
> const lower = {"lower"};

///////////////////////////////////////////////////////////////////////////////
/// \brief Matches a printable character.
///
/// The regex traits are used to determine which characters are printable.
/// To match any character that is not printable, use ~print.
///
/// \attention print is equivalent to /[[:print:]]/ in perl. ~print is equivalent
/// to /[[:^print:]]/ in perl.
proto::op_proxy<
    proto::unary_op<detail::posix_charset_placeholder, proto::noop_tag>
  , char const *
> const print = {"print"};

///////////////////////////////////////////////////////////////////////////////
/// \brief Matches a punctuation character.
///
/// The regex traits are used to determine which characters are punctuation.
/// To match any character that is not punctuation, use ~punct.
///
/// \attention punct is equivalent to /[[:punct:]]/ in perl. ~punct is equivalent
/// to /[[:^punct:]]/ in perl.
proto::op_proxy<
    proto::unary_op<detail::posix_charset_placeholder, proto::noop_tag>
  , char const *
> const punct = {"punct"};

///////////////////////////////////////////////////////////////////////////////
/// \brief Matches a space character.
///
/// The regex traits are used to determine which characters are space characters.
/// To match any character that is not white-space, use ~space.
///
/// \attention space is equivalent to /[[:space:]]/ in perl. ~space is equivalent
/// to /[[:^space:]]/ in perl.
proto::op_proxy<
    proto::unary_op<detail::posix_charset_placeholder, proto::noop_tag>
  , char const *
> const space = {"space"};

///////////////////////////////////////////////////////////////////////////////
/// \brief Matches an upper-case character.
///
/// The regex traits are used to determine which characters are upper-case.
/// To match any character that is not upper-case, use ~upper.
///
/// \attention upper is equivalent to /[[:upper:]]/ in perl. ~upper is equivalent
/// to /[[:^upper:]]/ in perl.
proto::op_proxy<
    proto::unary_op<detail::posix_charset_placeholder, proto::noop_tag>
  , char const *
> const upper = {"upper"};

///////////////////////////////////////////////////////////////////////////////
/// \brief Matches a hexadecimal digit character.
///
/// The regex traits are used to determine which characters are hex digits.
/// To match any character that is not a hex digit, use ~xdigit.
///
/// \attention xdigit is equivalent to /[[:xdigit:]]/ in perl. ~xdigit is equivalent
/// to /[[:^xdigit:]]/ in perl.
proto::op_proxy<
    proto::unary_op<detail::posix_charset_placeholder, proto::noop_tag>
  , char const *
> const xdigit = {"xdigit"};

///////////////////////////////////////////////////////////////////////////////
/// \brief Beginning of sequence assertion.
///
/// For the character sequence [begin, end), 'bos' matches the
/// zero-width sub-sequence [begin, begin).
proto::op_proxy<
    proto::unary_op<detail::assert_bos_matcher, proto::noop_tag>
> const bos = {};

///////////////////////////////////////////////////////////////////////////////
/// \brief End of sequence assertion. 
///
/// For the character sequence [begin, end),
/// 'eos' matches the zero-width sub-sequence [end, end).
///
/// \attention Unlike the perl end of sequence assertion \$, 'eos' will
/// not match at the position [end-1, end-1) if *(end-1) is '\\n'. To
/// get that behavior, use (!_n >> eos).
proto::op_proxy<
    proto::unary_op<detail::assert_eos_matcher, proto::noop_tag>
> const eos = {};

///////////////////////////////////////////////////////////////////////////////
/// \brief Beginning of line assertion. 
///
/// 'bol' matches the zero-width sub-sequence
/// immediately following a logical newline sequence. The regex traits
/// is used to determine what constitutes a logical newline sequence.
proto::op_proxy<
    proto::unary_op<detail::assert_bol_placeholder, proto::noop_tag>
> const bol = {};

///////////////////////////////////////////////////////////////////////////////
/// \brief End of line assertion.
///
/// 'eol' matches the zero-width sub-sequence
/// immediately preceeding a logical newline sequence. The regex traits
/// is used to determine what constitutes a logical newline sequence.
proto::op_proxy<
    proto::unary_op<detail::assert_eol_placeholder, proto::noop_tag>
> const eol = {};

///////////////////////////////////////////////////////////////////////////////
/// \brief Beginning of word assertion. 
///
/// 'bow' matches the zero-width sub-sequence
/// immediately following a non-word character and preceeding a word character.
/// The regex traits are used to determine what constitutes a word character.
proto::op_proxy<
    proto::unary_op<detail::assert_word_begin, proto::noop_tag>
> const bow = {};

///////////////////////////////////////////////////////////////////////////////
/// \brief End of word assertion. 
///
/// 'eow' matches the zero-width sub-sequence
/// immediately following a word character and preceeding a non-word character.
/// The regex traits are used to determine what constitutes a word character.
proto::op_proxy<
    proto::unary_op<detail::assert_word_end, proto::noop_tag>
> const eow = {};

///////////////////////////////////////////////////////////////////////////////
/// \brief Word boundary assertion.
///
/// '_b' matches the zero-width sub-sequence at the beginning or the end of a word.
/// It is equivalent to (bow | eow). The regex traits are used to determine what
/// constitutes a word character. To match a non-word boundary, use ~_b.
///
/// \attention _b is like \\b in perl. ~_b is like \\B in perl.
proto::op_proxy<
    proto::unary_op<detail::assert_word_boundary, proto::noop_tag>
> const _b = {};

///////////////////////////////////////////////////////////////////////////////
/// \brief Matches a word character.
///
/// '_w' matches a single word character. The regex traits are used to determine which
/// characters are word characters. Use ~_w to match a character that is not a word
/// character.
///
/// \attention _w is like \\w in perl. ~_w is like \\W in perl.
proto::op_proxy<
    proto::unary_op<detail::posix_charset_placeholder, proto::noop_tag>
  , char const *
> const _w = {"w"};